The speed of postharvest technology adoption in Tanzania: the role of social learning and agricultural extension services

This study examines the impacts of social learning and extension services on the time it takes to adopt an improved postharvest technology called Purdue Improved Crop Storage (PICS) bag in Tanzania. We utilized the doubly robust multivalued inverse probability weighted regression (MIPWRA) model in a survival treatment effects framework to estimate the impact. We also applied the Laplace regression model to evaluate the heterogeneous effects of the two information sources. Overall, results from the MIPWRA indicate that social learning and extension services reduce the time to adopt PICS bags by 51% and 49%, respectively. The results further show that the speed at which farmers adopted the technology was faster when using the two information sources jointly (61%) than individually. The results from the Laplace regression model also show that the impacts of social learning and extension services vary significantly across the time to adoption distribution. The marginal impacts of the two information sources are more meaningful for the households in the upper quantiles of the distribution, compared to the lower quantiles representing the early adopters. Designing policies that account for the complementarity of the two sources of information is essential to increasing the adoption of PICS bags in Tanzania.